The aim of this study was to identify differences between the individuals who committed homicide in BiH during the period from 01. January 1999. to 31. December 2009. in the evidently “violent” and “non-violent” way.

Two groups were investigated: 125 “violent” cases who committed homicide of an intimate partner (or filicide) and 125 “non-violent homicides” who killed people who were not members of the immediate family. Social-demographic and psychosocial characteristics of the two groups were investigated and compared. Instruments used in the study included: General Data List (GDL), test of personality characteristics (EPQ - Eizenck HJ), anxiety test (BAI - Beck AT), and emotion index profile (P.I.E - Plutchik R.).

Multivariant regressive analysis differentiates violent from non-violent murderers by predictors: education of father (R = 0.950, x2 = 35.10, df = 2, P < 0.001, OR = 0.125 (95%), Cl = 0.127–0.558), living in the place of the crime scene enviroment (R = 0.815, x2 = 46.75, df = 2, P < 0.001, OR = 0.105, CI = 0.110–0.217), gambling abuse and possible recidivism. Within the psycho-logical predictors domain, violent murderers are differentiated from the non-violent ones by psyhoticism (R = 0.750, df = 2, P < 0.001, OR = 0.710 (95%), CI = 0.290–1.770), destructiveness (R = 0.610, df = 2, P < 0.001, OR = 0.575 (95%), CI 0.970–1.435) incorporation in the PIE tests.

For the violent homicides, owing to personal and family history, presence of the microsocial model of violence transfer augmented by heredity. Important hereditary destructiveness was also found and stressed for the violent group. For the cases of nonviolent homicides several facts as poverty and long-term vulnerability were found to provoke accidental homicide.